An exemplary semiconductor wafer 100 is shown in FIG. 1A. The wafer 100 is formed by slicing a thin circular disk from a purified block of silicon. The wafer 100 may have a thickness ranging from 500 to 1000 microns. A typical wafer 100 may have a thickness of approximately 740 microns. An integrated circuit fabrication process may be used to form a plurality of semiconductor dies 101 upon the wafer 100. FIG. 1B is a cross sectional view of the wafer 100 and shows that the dies 101 are located near the front surface 103 of the wafer. After the dies 101 have been formed, a sawing process is typically used to separate the dies 101 from the wafer 100. Since the dies 101 are separated portions of the wafer 100, the dies 101 have the same thickness as the wafer 100 (e.g., 740 microns).
Each die 101 may be mounted within a package 150 to form a semiconductor chip 200, as shown in FIG. 2. The package 150, which is designed to protect the die 101 and to coupled a plurality of leads 125 to the die 101, is only compatible with dies 101 having a specified range of thickness. For example, a commercial die package may be designed to accommodate dies having a thickness of approximately 305 microns. It is often desirable to reduce the thickness of the semiconductor package. One method of reducing the thickness of a package is to use a thinner package, which often requires the use of thinner dies 101. For example, some packages may only be compatible with significantly thinner (e.g., 100 micron thick) dies 101. Thus, both conventional and thin profile packages often require dies 101 which are much thinner than most wafers 100.
The procedure to reduce the thickness of the dies 101 so that they are compatible with a given package design is known as backgrinding. This procedure takes advantage of the fact that the dies 101 are formed near the front surface 103 of the wafer. Thus, the back surface 104 of the wafer may be ground down to reduce the thickness of the wafer 100, and the dies 101 formed thereon, if the mechanical stress associated with backgrinding can be controlled to avoid fracturing the wafer 100 or damaging the dies 101. FIG. 3 is a block diagram of a backgrinder 402, which includes a chucktable 300 and grinding wheel 310. The chucktable 300 is used to flatly support the wafer 100 as it is backgrinded by the grinding wheel 310. The flat support offered by the chucltable 300 distributes stress induced by the grinding wheel 310, thereby reducing the chance of wafer fracture. Additionally, referring also to FIG. 5, a layer of protective tape 320 is attached to the front surface 103 of the wafer 100 by a tape applicator 400. Thus, the layer of protective tape 320 lies between the front surface 103 of the wafer 100 and the chucktable 300, thereby protecting the dies 101 and further absorbing mechanical stress.
However, as illustrated by FIG. 4, when excess tape 325 extends significantly beyond the perimeter of the wafer 100, that excess tape 325, also known as tape bur 325, may become trapped and folded between the chucktable 300 and the wafer 100. The tape bur 325 can prevent the chucktable 300 from flatly supporting the wafer 100, and increase the possibility of uneven back side grinding and possible wafer fracture.
In order to minimize this problem, the wafers 100 are processed by a tape cutter 401 before they are backgrinded. The tape cutter 401 is responsible for trimming the protective tape 320 at or near the perimeter of the wafer 100, thereby removing any tape bur 325. However, the tape cutter 401 cannot consistently guarantee that each processed wafer 100 is free of tape bur 325 because the tape cutter 401 is susceptible to several malfunctions. Thus, under certain circumstances, a tape cutter 401 may output a wafer 100 with tape bur 325 for subsequent processing by the backgrinder 402. Accordingly, there is a need for an apparatus and a method to prevent a tape cutter from outputting a wafer with tape bur for subsequent processing by a backgrinder.